Effectiveness of Deep Learning Long Short-Term Memory Network for Stock Price Prediction on Graphics Processing Unit

The stock price is an important barometer of a country's economic prosperity. As a result, detecting and describing the precise changes in the stock price is extremely valuable. However, because the stock market's complicated and uncertain behavior makes precise forecasting impossible, strong predicting models are very beneficial for the financial decision-making processes of investors. For many experts and analysts, projecting stock prices has proven to be a difficult undertaking. Indeed, investors are keenly engaged in the field of stock price forecasting research. Numerous stockholders are concerned in the future direction of the stock price in order to make a wise and successful investment. Effective stock market prediction systems assist traders, investors, and analysts by giving helpful information such as the stock market's future direction. A unique LSTM model is utilized to predict stock market prices in this study. In order to shorten training and testing time, the suggested model experimental analysis was performed on Graphics Processing Unit (GPU). Data denoising and normalization were used in the data preprocessing phase. The suggested model was tested using experimental datasets from the S&P 500, NYSE, Nasdaq, and Forbes, and the findings were compared to those of other models. When compared to existing approaches, the experimental findings clearly show improvements in MAE, RMSE, and MAPE with competitive performance.